Background
Quantum computers store information in quantum bits, which are fragile and make errors easily. Quantum error correction uses extra quantum bits to detect and fix those errors, and researching it requires classical computers to decode the error signals quickly. Making quantum processors work with ordinary supercomputers is therefore a practical problem, not only a hardware one.
What happened
On November 16, 2026, NVIDIA announced plans for collaborations with IBM and the US Department of Energy's Fermi National Accelerator Laboratory (Fermilab).[1] The work covers modular quantum computing architectures and the integration of quantum systems with supercomputers.[1] The partners plan a modular quantum-classical codesign platform for quantum error-correction and applications research.[1]
Key details
NVIDIA DGX B200 systems will be colocated with IBM quantum computers, including those powered by IBM Quantum Nighthawk.[1] Colocation means the classical systems and the quantum computers sit in the same place, so they can exchange data with little delay.
NVIDIA also released TCBP-JOBS, a decoding algorithm in CUDA-Q. NVIDIA says it delivers up to a 13,000x speedup for quantum error-correction research.[1] The company said the algorithm cuts the computation time for researching new quantum computing designs from 12 weeks to about 45 minutes.[1]
NVIDIA is also expanding Slurm, a widely used job scheduler for supercomputers, to schedule quantum processors in supercomputing workflows. IBM is contributing integration through its Quantum Resource Management Interface.[1]
Quantum processors and classical supercomputers run on different kinds of hardware and software, so connecting them requires shared scheduling and data interfaces. Slurm and the Quantum Resource Management Interface are shared layers of this kind, and they let a job queue treat a quantum processor as one more resource in a workflow.
Impact
The collaboration tries to connect quantum processors to the same scheduling and computing infrastructure that researchers already use for classical simulation. The 12-week-to-45-minute figure is NVIDIA's own measurement, and the partners describe the platform as a plan, so the work is still in progress.
Sources
- 1
Cite this page: NVIDIA, IBM and Fermilab announce quantum-classical codesign collaboration. TellWorld, https://www.tellworld.net/news/nvidia-ibm-fermilab-quantum-2026. Updated 2026-11-18